A novel class of AMPA receptor allosteric modulators. Part 1: Design, synthesis, and SAR of 3-aryl-4-cyano-5-substituted-heteroaryl-2-carboxylic acid derivatives

Article abstract of DOI:10.1016/j.bmcl.2006.07.035

The synthesis and initial SAR studies of novel, highly potent positive allosteric modulators of AMPA receptors based on 3-(4-tert-butylphenyl)-4-cyano-5-methylsulfanyl-thiophene-2-carboxylic acid (6a) are described. SAR studies at the thioether moiety ind

Full text of DOI:10.1016/j.bmcl.2006.07.035

Bioorganic & Medicinal Chemistry Letters 16 (2006) 5057–5061
A novel class of AMPA receptor allosteric modulators.
Part 1: Design, synthesis, and SAR of 3-aryl-4-cyano-5-
substituted-heteroaryl-2-carboxylic acid derivatives
Maria-Carmen Fernandez,a,* Ana Castano,a Esteban Dominguez,a Ana Escribano,a
˜
Delu Jiang,b Alma Jimenez,a Eric Hong,b William J. Hornback,b
Eric S. Nisenbaum,b Nancy Rankl,b Eric Tromiczak,b Grant Vaught,b
Hamideh Zarrinmayehb and Dennis M. Zimmermanb
aLilly S.A., Avenida de la Industria, 30, 28108 Alcobendas, Madrid, Spain
bEli Lilly and Co., Lilly Research Laboratories, Indianapolis, IN 46285, USA
Received 11 May 2006; revised 12 July 2006; accepted 13 July 2006
Available online 1 August 2006
Abstract—The synthesis and initial SAR studies of novel, highly potent positive allosteric modulators of AMPA receptors based on
3-(4-tert-butylphenyl)-4-cyano-5-methylsulfanyl-thiophene-2-carboxylic acid (6a) are described. SAR studies at the thioether moiety
indicated that substitution at this position was mandatory and better potency was achieved with small groups.
Ó 2006 Elsevier Ltd. All rights reserved.
L-Glutamate is a major excitatory neurotransmitter in
mammalian central nervous system (CNS). Three iono-
tropic glutamate receptor subtypes have been identified
based on differences in their molecular biology and sen-
sitivity to the selective agonists N-methyl-D-aspartate
deficits associated with a variety of neurological and
psychiatric disorders.5
Clinical and experimental data suggest that positive
modulation of AMPA receptors may be therapeutically
effective in the treatment of cognitive deficits.6 Several
studies have identified compounds that enhance ion
influx through AMPA receptors by positive allosteric
modulation. Some examples are pyrrolidones such as
aniracetam7a–c (1, Fig. 1), benzothiadiazides such as
cyclothiazide (2),7d,e benzamides such as CX-516 (3),
or 4-[2-(phenylsulfonylamino)thio]-difluoro-phenoxy-
acetamide (4),7f and a family of sulfonamides (5a and
5b) previously reported by our group.7g
(NMDA),
a-amino-3-hydroxy-5-methyl-4-isoxazole-
propionic acid (AMPA) and kainate (KA).1 The AMPA
receptor family includes four different subunits named
GluR1-4. Functional AMPA receptors are most likely
tetrameric ion channels formed by the assembly of one
or more of the GluR subunits, yielding homomeric or
heteromeric receptors.2 Additional complexity among
AMPA receptors is conferred by alternative splicing of
RNA for each subunit giving rise to flip and flop vari-
ants.3 Glutamate binding to AMPA receptors initiates
conformational changes in the channel gate that lead
to increase conductance of sodium and calcium ions into
cells and subsequent membrane depolarization.4
Electrophysiological experiments have demonstrated
that these biarylsulfonamides are potent modulators of
native AMPA receptors enhancing glutamate evoked
currents through AMPA channels of acutely isolated
pyramidal neurons.8a Moreover, in vivo tests on spinal
and hippocampal neurons in rats showed that 5a and
5b cross the blood–brain barrier and modulate the sen-
sitivity of central AMPA receptors enhancing synaptic
excitation.8b
Evidence indicates that decreases in glutamatergic neu-
rotransmission in the CNS may contribute to cognitive
Keywords: AMPA receptor allosteric modulators; Heteroaryl carbox-
ylic acids.
Toward the goal of developing new AMPA potenti-
ators, a high throughput screen of compounds was
*
Corresponding author. Tel.: +34 91 663 3437; fax: +34 91 663
3411; e-mail: fernandez_maria-carmen@lilly.com
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.07.035

5058
M.-C. Fernandez et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5057–5061
intermediates were further converted to the thiophenyl
(13a and 14),11 N-Me pyrrole (15 and 16), and furan
(17) ethyl carboxylates by cyclocondensation with ethyl-
thioglycolate, sarcosine ethyl ester hydrochloride and
bromoethyl acetate, respectively, under basic conditions.
Carboxylic acids 6a, 18–21 were prepared from esters
13–17 by hydrolysis under alkaline conditions.
Intermediate 13a was further elaborated to give a wide
range of derivatives summarized in Scheme 2. Thioph-
enes 6c–f were obtained by coupling of primary thiols
with 13a in toluene, using palladium chemistry. Oxida-
tion of methylthio group with MCPBA followed by
ester hydrolysis afforded the sulfoxide 6g, while excess
of the same oxidant yielded the sulfone 13h. This inter-
mediate was used to prepare the amino ester precursors
of 6i–j by nucleophilic substitution of the corresponding
amines. Moreover, 13h reacted with diethylzinc in dry
dichloromethane to provide the ethylthiophenyl carbox-
ylic ester intermediate that after ester hydrolysis affor-
ded 6k. Compound 6b was obtained by reduction of
13 h with Na–Hg amalgam in presence of base followed
by hydrolysis. Alternatively, this compound could be
prepared by reaction of 13a with excess of sodium boro-
hydride in ethanol and subsequent ester hydrolysis.
Figure 1. Structures of AMPA receptor allosteric modulators.
conducted using human cloned homomeric AMPA
receptors. The screen identified thiophene 6a9 (Fig. 1)
as a novel AMPA receptor modulator lead. Further
structure–activity-relationship (SAR) studies led to the
discovery of a new family of AMPA potentiators. In
the present communication, we describe part of these
SAR studies, focusing our attention on the identification
of the optimal core ring and the substitution at
thiomethyl moiety in thiophene ring.
Compound 6l was prepared using the synthetic
approach described in Scheme 3. Condensation of (tri-
fluoromethyl)ethylene 2212 with ethyl 2-mercaptoacetate
in ethanol afforded the aminothiophene 23. This inter-
mediate reacted with isoamylnitrite and diiodomethane
to provide the iodo derivative 24. Then, Suzuki coupling
of 24 with tert-butylphenyl boronic acid under standard
conditions and subsequent ester hydrolysis provided
thiophene derivative 6l.
The synthetic pathway for preparing compounds 6a, 18–
21 is shown in Scheme 1.
Methylsulfanyl acrylonitriles 11 and 12 were obtained
by deprotonation of the corresponding propionitriles,
previously prepared under standard conditions,
followed by condensation with carbon disulfide and
subsequent quenching with methyl iodide.10 These
All compounds were dissolved in DMSO and tested for
activity in calcium-imaging assay using stably transfec-
ted HEK-293 cells expressing homomeric GluR2 and
GluR4 receptors in both flip and flop splice variant
forms to provide information on the relative potency
and splice variant selectivity. Changes in intracellular
calcium concentration produced by application of gluta-
mate (100 lM) alone and in response to co-application
with compounds (0.1–3 lM) were measured using a
Fluorometric Imagin Plate Reader (FLIPR) technology.13
The activity of compounds at various concentrations
was expressed as a percentage of the response evoked
by a saturating concentration (100 lM) of cyclothiazide
(2), and EC50 values were calculated. These data are
shown in Tables 1 and 2.
The lead compound 6a (Fig. 1) showed potentiator activ-
ity with EC50 values of 214 nM and 1558 nM at GluR4
flip and GluR4 flop receptors, and 97 nM and 309 nM
at GluR2 flip and GluR2 flop receptors, respectively. Flip
splice variant preferences were observed in both GluR2
and GluR4 receptors, although higher potency was
exhibited in GluR2 receptors. Comparing thiophene 6a
with the previously reported sulfonamides 5a and 5b in
GluR4 flip receptors; thiophene 6a was equipotent with
5b and 2-fold more potent than 5a. Moreover,
cyclothiazide 2 was 17-fold less active than 6a.
Scheme 1. Synthesis of compounds 6a, 18–21. Reagents and
conditions: (a) i—SOCl2, 50 °C, 1 h; ii—CNCH2CO2H, n-BuLi,
THF, À78 °C to rt, 2h (50–64%); (b) i—CS2, DMSO, 15 °C to rt,
2h; ii—MeI, rt, 1 h (42–56%); (c) ethylthioglycolate, Et3N, EtOH,
reflux, 15 min (X = S); CH3NHCH2CO2EtÆHCl, Et3N, EtOH, 0.5–2 h
(X = NMe); Bromoethyl acetate, LHMDS, THF, À78 °C to rt, 6 h
(X = O) (50–73%); (d) 2 M NaOH/EtOH 1:1 or 2.5 M LiOH/THF/
MeOH 3:2:1, rt, overnight, then 1 M HCl (30–95%).

M.-C. Fernandez et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5057–5061
5059
Scheme 2. Synthesis of compounds 6b–k. Reagents and conditions: (a) RSH, t-BuOK, (R)-(+)-BINAP, Pd(PPh3)4, toluene, 90 °C, (75–90%); (b)
MCPBA (2.0 equiv), DCM, 0 °C to rt (47%); (c) 2 M NaOH/EtOH 1:1, rt, overnight, then 1 M HCl (35–80%); (d) Me2NH (2 N) in THF, rt (55%); (e)
pyrrolidine, rt (90%); (f) Et2Zn, DCM, rt, 5 days; (g) Na–Hg, Na2HPO4, MeOH, rt, 48 h; (h) NaBH4, EtOH, 0 °C (30%); (i) MCPBA (1.0 equiv),
DCM, 0 °C to rt (53%).
Further details of SAR results in this region will be
reported in further communications.
Once we confirmed that thiophene was one of the most
potent five-membered heterocycles, we turned our atten-
tion to evaluating the effects of thioether replacements
by other substituents. We started varying the alkyl
group in thioether moiety (6a, 6c–f, Table 2). All modi-
fications in this region maintained flip selectivity in
GluR4 receptor, but potency was more dependent on
the steric hindrance of the substituents. In this way,
activity decreased as larger groups were introduced
(compounds 6d–f). However, substitution at this posi-
tion is mandatory since the unsubstituted thiophene 6b
did not show activity in any receptor at tested
concentrations.
Scheme 3. Synthesis of compound 6l. Reagents and conditions: (a)
ethyl 2-mercaptoacetate, KOAc, EtOH, 60–70 °C, 30 min (31%); (b)
CH2I2, CH3CN, isoamylnitrite, 35–65 °C, 10 min (81%); (c) i—tert-
butyl phenylboronic acid, Pd(PPh3)4, Na2CO3, ethyleneglycol, 80 °C,
overnight; ii—2 M NaOH/EtOH 1:1, rt, overnight, then 1 M HCl
(25%, two steps).
We next explored the effect of the heterocycle as modu-
lator of AMPA receptors. For these studies, the carbox-
ylic acid, cyano and aryl groups were kept intact while
the thiophene moiety was replaced by furan and
N-methypyrrole. As shown in Table 1, N-methylpyrrole
19a was slightly less potent than 6a at both flip and flop
isoforms of GluR4 receptors but 7-fold less potent at
GluR2 flip receptors. However, replacement of sulfur
by oxygen (compound 21) was not well tolerated reduc-
ing the potency 9-fold in GluR4 flip. Based on these
results, we chose the thiophene and pyrrole cores and
replaced the tert-butyl by a phenyl group (compounds
18 and 20 respectively) in order to avoid the potential
metabolism issues related to tert-butyl group. In this
case, phenyl derivative 20 increased the potency, slightly
in case of GluR4 and 4-fold in GluR2 flip receptors
compared to the tert-butyl analogue 19a. However, this
effect was not observed in thiophene derivative 18.
We also investigated whether groups with opposite elec-
tronic or polar features could replace thiomethyl
moiety. Starting from the readily accessible oxidation
products derived from 6a, sulfoxide 6g and sulfone 6h,
we observed that in both cases, these electron-withdraw-
ing substituents decreased the activity at GluR4 flip
receptors and did not show activity in GluR4 flop.
However, trifluoromethyl derivative 6l shown similar
potency to 6a in GluR4 (EC50 = 273 nM) and GluR2
(EC50 = 78 nM) flip receptors, but was 3-fold more
potent in GluR4 flop (EC50 = 488 nM). The replace-
ment of thiomethyl by ethyl group (entry 6k) was well
tolerated and exhibited similar potency and selectivity
to 6a. However, amino substitution (entries 6i and 6j)
reduced the activity in tested receptors following the

5060
M.-C. Fernandez et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5057–5061
Table 1. EC50 values of selected novel AMPA receptor allosteric modulators
Compound
X
R
EC50 (nM)
GluR4-flip
GluR4-flop
GluR2-flip
GluR2-flop
2a
—
—
—
—
3800
660
420
214
795
554
394
1979
NTb
NT
NT
NT
559
97
NT
NT
1905
309
NT
663
474
NT
5aa
5ba
6a
—
S
—
1061
1558
>3000
1423
1301
>3000
t-Bu
Ph
18
S
NT
684
174
NT
19a
20
NMe
NMe
O
t-Bu
Ph
21
t-Bu
a See Figure 1 for structures of 2, 5a, and 5b.
b NT, not tested.
Table 2. EC50 values of selected novel AMPA receptor allosteric modulators
Compound
R
EC50 (nM)
GluR4-flip
GluR4-flop
GluR2-flip
97
GluR2-flop
6a
6b
6c
6d
6e
6f
MeS
H
214
>3000
440
1558
309
>3000
>3000
>3000
>3000
>3000
>3000
>3000
>3000
>3000
NTa
NT
>3000
NT
NT
EtS
n-PrS
i-PrS
PhS
2892
1873
>3000
489
NT
NT
NT
NT
6g
6h
6i
SOMe
SO2Me
NMe2
NT
333
NT
629
474
1808
2854
176
6j
1009
>3000
1137
2433
6k
6l
Et
CF3
183
273
1221
488
47
78
216
231
a NT, not tested.
same trend as thioether derivatives (6a, 6c–f). The
more hindered 5-pyrrolidinyl thiophene 6j became less
potent than the dimethyl amino analogue 6i in both
GluR2 and GluR4 receptors.
studies demonstrated that the N-methylpyrrole deriva-
tive 19a showed similar potency to the thiophene lead
6a at GluR4 receptors and was only slightly less active
at GluR2 receptors. SAR trends at the thioether moiety
were achieved and better potency was found with
small groups independently of the electronic features.
However, substitution at this position was mandatory.
Flip splice variant selectivity was also predominant in
both series.
We also observed that carboxylic acid played a crucial
role in terms of activity. While 6a was very potent in both
GluR2 and GluR4 receptors (Table 2), the ester precursor
13a did not show activity in any receptor
(EC50 > 3000 nM in GluR4 and GluR2 flip and flop) at
tested concentrations. This result suggested that carbox-
ylic acid could be establishing a hydrogen bond interac-
tion with the receptor. In order to test this hypothesis,
we explored more thoroughly this domain. Those results
will be reported in further communications.
Acknowledgments
´
The authors gratefully acknowledge J. L. Roldan and
J. C. de la Rosa and Prf. J. C. Carretero from
´
Universidad Autonoma de Madrid for their contribu-
In summary, we have designed, synthesized, and evalu-
ated a novel class of AMPA potentiators. Initial SAR
tions to this paper and to the Analytical Technologies
group at Lilly for the analytical work.

M.-C. Fernandez et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5057–5061
5061
Gupta, A.; Zimmerman, D. M.; Bleakman, D. Neuro-
pharmacology 2001, 40, 976.
References and notes
8. (a) Baumbarger, P.; Muhlhauser, M.; Yang, C. R.;
Nisenbaum, E. S. Neuropharmacology 2001, 40, 992,
and references cited herein; (b) Vandergriff, J.; Huff,
K.; Bond, A.; Lodge, D. Neuropharmacology 2001, 40,
1003.
9. Similar structures have been recently disclosed by Nich-
olson et al as one of the small molecules that enhance the
catalytic efficiency of HLA-DM, for further information,
see: Nicholson, M.; Moradi, B.; Seth, N. P.; Xing, X.;
Cuny, G.; Stein, R. L.; Wucherpfenning, K. W. J.
Immunol. 2006, 176, 4208.
1. Hollmann, M.; Heinemann, S. Annu. Rev. Neurosci. 1994,
17, 31.
2. (a) Dingledine, R.; Borges, K.; Bowie, D.; Traynelis, S. F.
Pharmacol. Rev. 1999, 51, 7; (b) Schoepp, D. D.; Jane, D.
E.; Monn, J. A. Neuropharmacology 1999, 38, 1431.
3. (a) Seeburg, P. H.; Higuchi, M.; Sprengel, R. Brain Res.
1998, 26, 217; (b) Hume, R. I.; Dingledine, R.; Heine-
mann, S. F. Science 1991, 253, 1028.
4. (a) Fletcher, E. J.; Lodge, D. Pharmacol. Ther. 1996, 70, 1,
65; (b) Bleakman, D.; Lodge, D. Neuropharmacology
1998, 37, 1187.
10. Gompper, R.; Kutter, E. Angew. Chem. 1962, 7, 251.
11. Augustin, M.; Rudorf, W.-D.; Schmidt, U. Tetrahedron
1976, 32, 3055.
5. Yamada, K. A. Expert Opin. Investig. Drugs 2000, 9, 765.
6. (a) Dimond, S. J.; Scammell, R. E.; Pryce, I. C.; Huws, D.;
Gray, C. Psychopharmacology 1979, 64, 341; (b) Ingvar,
M.; Ambros-Ingerson, J.; Davis, M.; Granger, R.; Kessler,
M.; Rogers, G. A.; Schehr, R. S.; Lynch, G. Exp. Neurol.
1997, 146, 533; (c) Staubl, U.; Perez, Y.; Xu, F. Proc. Natl.
Acad. Sci. U.S.A. 1994, 91, 777; (d) Yamada, K. A.
Neurobiol. Dis. 1998, 5, 67.
7. (a) Ito, I.; Tanabe, S.; Kohda, A.; Sugiyama, H. J. Physiol.
1990, 424, 10940; (b) Staubli, U.; Kessler, M.; Lynch, G.
Psychobiology 1990, 18, 377; (c) Arai, A.; Kessler, M.;
Xiao, P.; Ambros-Ingerson, J.; Rogers, G.; Lynch, G.
Brain Res. 1994, 638, 343; (d) Yamada, K. A.; Rothman,
S. M. J. Physiol. 1992, 458, 409; (e) Desos, P.; Serkiz, B.;
Morain, P.; Lepagnol, J.; Cordi, A. Bioorg. Med. Chem.
Lett. 1996, 6, 3003; (f) Sekiguchi, M.; Fleck, M. W.;
Mayer, M. L.; Takeo, J.; Chiba, Y.; Yamashita, S.; Wada,
K. J. Neurosci. 1997, 17, 5760; (g) Miu, P.; Jarvie, K. R.;
Radhakrishnan, V.; Gates, M. R.; Ogden, A.; Ornstein, P.
L.; Zarrinmayeh, H.; Ho, K.; Peters, D.; Grabell, J.;
12. Middleton, J. J. Fluorine Chem. 1982, 20, 397.
13. Typical protocol for this assay: 96-Well plates containing
confluent monolayers of HEK-293 cell stably expressing
human AMPA receptors were prepared. Cells were
incubated in buffer solution (5.5 mM glucose, 136.9 mM
NaCl, 0.8 mM MgSO4Æ7H2O, 5.4 mM KCl, 5 mM
CaCl2Æ2H2O, 0.44 mM KH2PO4, 4.2 mM NaHCO3,
0.34 mM Na2HPO4Æ7H2O, and 20 mM HEPES) contain-
ing 8 lM Fluo3-AM dye (obtained from Molecular
Probes Inc., Eugene, OR) for 60 min. Measurements were
made using a FLIPR that indicates changes in fluorescence
upon influx of calcium into cells upon stimulation by
glutamate (100 lM) in the presence of cyclothiazide
(100 lM) or compound. Each compound was run in
duplicate 10 point curves and each concentration and
replicate were in a single well on the same plate on the
same cells.